| Abstract |
Subsidence, and its effects, from conventional geothermal operations in New Zealand have been well documented and closely studied for many decades. Examples are presented here of local subsidence anomalies that show associations with relatively-shallow, anomalously-compressible, porous formations, weakened by hydrothermal alteration from boiling fluids that passed through shallow outflow structures. Pressure decline, originating from deep production, but diffusing slowly into shallow aquifers and aquicludes, has previously been attributed as the principal cause. Anomalous deformation is found to be the product of subsurface changes in effective stress (either pressure or temperature in origin), acting on a thick sequence of clays, or fractured rocks, which exhibit anomalous geo-mechanical properties. Consideration of the transition between brittle failure and ductile behaviour across a range of temperatures and rock types is also needed. Settlement can increase over time due to non-linear stress-strain relationships such as clay yielding. This paper reviews New Zealand geothermal case studies from Wairakei, Tauhara, Ohaaki and Kawerau, where applications were recently granted for resource consent renewal or development expansion. These efforts have stimulated additional studies of observed changes in subsidence rates, corresponding horizontal deformation, inferred deformation mechanisms, model predictions and possible mitigation options. Mechanisms involving subsurface temperature change and chemical alteration, as well as transient pressures, are sometimes implicated. Transient tectonic creep, ‘shake-induced’ subsidence and groundwater level fluctuations acting on buried deposits of unconsolidated alluvium, are also plausible mechanisms for fluctuating deformation rates of natural origin in some settings (for example, Kawerau). To properly simulate the deformation processes, fully inter-coupled Thermal-Hydraulic-Mechanical-Chemical modelling would be preferred, but history matching suffers from a plethora of variables and a shortage of good subsurface data. Fundamental rock properties used in traditional reservoir simulation, such as permeability, porosity and stress state, which are usually treated as constant parameters in history matching and subsequent scenario predictions, turn out to be significant variables in deformation modelling. Alternative and more pragmatic modelling approaches, that simplify the geothermal subsidence process, have proven to be reasonably successful where rate changes are smoothly varying, but predictions retain significant uncertainty, particularly where non-geothermal mechanisms are important, and where the hydro-geological properties of shallow layers are poorly represented in the simple models. However, adaptive mitigation options for adverse effects of subsidence, using comprehensive monitoring, are generally accepted. These are usually expressed in terms of targeted injection management to control pressure and temperature. |